Beliefs related to adherence to oral antidiabetic treatment according to the Theory of Planned Behavior1

OBJECTIVE: to identify salient behavioral, normative, control and self-efficacy beliefs related to the behavior of adherence to oral antidiabetic agents, using the Theory of Planned Behavior. METHOD: cross-sectional, exploratory study with 17 diabetic patients in chronic use of oral antidiabetic medication and in outpatient follow-up. Individual interviews were recorded, transcribed and content-analyzed using pre-established categories. RESULTS: behavioral beliefs concerning advantages and disadvantages of adhering to medication emerged, such as the possibility of avoiding complications from diabetes, preventing or delaying the use of insulin, and a perception of side effects. The children of patients and physicians are seen as important social references who influence medication adherence. The factors that facilitate adherence include access to free-of-cost medication and taking medications associated with temporal markers. On the other hand, a complex therapeutic regimen was considered a factor that hinders adherence. Understanding how to use medication and forgetfulness impact the perception of patients regarding their ability to adhere to oral antidiabetic agents. CONCLUSION: medication adherence is a complex behavior permeated by behavioral, normative, control and self-efficacy beliefs that should be taken into account when assessing determinants of behavior.


Introduction
The use of non-pharmacological and pharmacological treatment is recommended to control and facilitate the management of Diabetes Mellitus (DM). Hence, special attention has been paid to factors related to medication adherence, since patients with chronic diseases, the emergence of which are asymptomatic, are more likely not to adhere to medication (1) . It is known that despite considerable technological advancements achieved in regard to the diagnosis and treatment of DM patients, a large number of people do not adhere to their recommended treatment (2)(3) .
In general, medication adherence is seen as a measure in which patients follow instructions for the treatments prescribed (4) . The decision to take medication or not is an empirical-rational strategy patients use to express their attempts to deal with the disease. It is important to acknowledge and not to underestimate the patients' decision-making abilities, identifying their beliefs and helping them adopt appropriate behaviors.
The Theory of Planned Behavior (TPB) (5) , one of the main models used to study health-related behaviors, assumes that beliefs impact predictive factors of intention (motivation) -immediate mediator of behavior. Hence, identifying beliefs regarding medication adherence is key to understanding patients' self-care actions because beliefs represent the opinions of patients, based on their knowledge or experiences, in regard to health orientation, influencing acceptance or rejection of the prescribed therapy (6) .
Even though TPB is widely used in studies seeking understanding various health behaviors, its application in studies addressing behaviors related to the prevention and/or control of DM is incipient. Among DM patients, there are studies addressing regular exercise (7) , insulin administration (8) , intake of low saturated-fat foods (9) , and adherence to oral antidiabetic agents (10) , however, no studies have been conducted with a Brazilian population with DM.
This study's aim was to identify and analyze salient beliefs -behavioral, normative, control and self-efficacy, related to adherence to oral antidiabetic agents through the application of TPB. Identifying these beliefs using TPB is an essential step to support the development of a measurement instrument to identify factors that determine adherence to oral antidiabetic agents. Understanding these determinants can aid the development of effective interventions to promote adherence to these medications.

Method
Cross-sectional, exploratory study with a qualitative and quantitative approach guided by the TPB theoretical framework.

Theoretical framework
This motivational model was derived from cognitive-social theories in which behavior is determined by intention (motivation) to act (effectively perform a given behavior) and by one's perception of control over behavior. Intention, the immediate antecedent of behavior, is determined by three factors: Attitude -an individual's assessment in regard to the probable or expected outcomes of a behavior; Subjective Norm -perceived social pressure, that is, the individual's perception concerning the opinion of social references in regard to a given behavior; and Perceived Behavioral Control -the individual's perception in regard to his/her control over behavior (5,11) . Each of the three determinants of Intention is formed by their respective beliefs. Behavioral beliefs refer to the assessment of results of behavior and produce a favorable or unfavorable attitude toward it. Normative beliefs refer to the perception of social references concerning a given behavior and result in a subjective norm. Control beliefs are created with the presence of perceived factors that may either facilitate or hinder a given behavior and influence perceived behavioral control (5,11) . The model recommends that, when identifying beliefs, salient beliefs, i.e., those that come first to mind when an interviewee is asked open questions about behavior, should be taken into account (12) .
Even though the literature indicates the strong predictive power of TPB in determining behavioral intention in regard to a complex range of behaviors, researchers (13) have defended the inclusion of other variables as being able to improve the explanation of variability of motivation and its predictive power for behavior. The selection of these variables depends on their relevance to understanding a desired behavior and the characteristics of the studied behavior. In regard to medication adherence, the variable Selfefficacy was considered relevant and refers to an individual's confidence in his/her ability to perform certain behaviors that influence the events affecting his/her life (14) .

Study setting
This study was conducted in a general adult outpatient clinic of a large university hospital in the interior of the state of São Paulo. Individuals whose administration of medication was managed by a caregiver, who used insulin, had been hospitalized or undergone surgery in the last 30 days, or showed impaired understanding or communication abilities, were excluded. A convenience sample was used. All the patients who met the inclusion criteria were consecutively added into the study until data saturation was achieved, i.e., when no new information was obtained and reports became redundant (15) .

Data collection
Data were obtained from March to April 2012 through individual semi-structured interviews, under a free-response format intended to identify salient beliefs (11) . The interviews were conducted and recorded in a private environment and later transcribed verbatim. normative (five), control (six) and self-efficacy beliefs (two). Behavior was defined as: "to take medication for diabetes treatment exactly as prescribed by my physician over the next two months", which includes target, action, context and time, in accordance with the TPB (11) premises.

Data analysis
The answers were submitted to content analysis based on the recommendations of the TBP theoretical model (11,18) . The following stages were followed: i) exploration of material in order to categorize it using categories pre-established by TPB, i.e., behavioral, normative, control and self-efficacy beliefs concerning adherence to oral antidiabetic medication; ii) the answers were grouped into each category according to themes or subcategories that were extracted from the participants' reports. The subcategories that emerged from the reports were submitted Rev. Latino-Am. Enfermagem 2014 July-Aug.;22(4):529-37.
to inter-observer analysis to ensure reliability of assessment (two researchers who were experts in the application of TPB), with an agreement level of 95% between the judges, and iii) the frequency of subcategories were calculated and modal/more frequent beliefs were highlighted. To identify which beliefs would be included, we used one of the criteria proposed by TPB, in which beliefs that exceed a given frequency are considered modal. All the beliefs mentioned in this study by at least 10% of the sample were included.

Ethical aspects
The Project was approved by the Institutional Review Board at a university in the interior of São Paulo, Brazil (Process No. 6.608/2012).

Results
The sample was composed of 17 individuals, most of whom were women (64.7%), aged 59.8 years old on average, with 3.9 years of schooling on average, and lived with other people (88.2%). Family income was 2.5 times the minimum wage on average (  Rev. Latino-Am. Enfermagem 2014 July-Aug.;22(4):529-37.

Normative Beliefs %
When taking medication for diabetes treatment exactly as prescribed, I would be approved by… My children 70.6 My physician 58.8 My spouse 52.9 The nursing staff 17.6 When taking medication for diabetes treatment exactly as prescribed, I would not be approved by… Acquaintances who are diabetic and do not adhere to treatment 58.8 Table 3 -Normative beliefs concerning the behavior of adhere to oral antidiabetic agents. Campinas, SP, Brazil, 2012   (Table 4).

Control Beliefs %
These make it easier to take medication for diabetes treatment exactly as prescribed… Access medication free-of-cost 70.6 Associate time of medication with temporal markers 29.4 Bring the pills whenever leaving home 29.4 Being able to differentiate pills by color, form and size 17.7 Having a routine and control over daily activities 11.8 This makes it harder to take medication for diabetes treatment exactly as prescribed… Having to take medication more than once a day 23.5 the design of effective interventions to promote this behavior among DM patients (19) .
Health professionals face the challenge of discovering the cause of hyperglycemia when dealing with patients whose DM is poorly managed; that is, the professional has to determine whether the poor management is related to treatment nonadherence or occurs despite the correct use of medications. Since patients may be more willing to report their negative beliefs concerning medication treatment than admit to low adherence, asking for their beliefs may enable the identification of patients with a greater likelihood of nonadherence to treatment.
In regard to behavioral beliefs, we highlight the impact of perceptions of reactions attributed to oral antidiabetic agents, which may lead to nonadherence.
This result was also found in previous studies (19)(20) .
Hypoglycemic episodes are also disadvantageous to adherence, responsible for irritation, sickness, and weakness, which contribute to the poor glycemic control of DM (21) . Fear of these events increases stress associated with DM, with an important impact on the management of the disease and metabolic control (22) . In regard to control beliefs, the factor that hindered or impeded medication adherence was the need to take medications more than once a day. It shows that the complexity of the therapeutic scheme is a factor that does not favor the correct use of medications. The factors that favored or facilitated adherence behavior were having access to free-of-cost medication, following a routine and having control over daily activities, taking medication associated with temporal markers, carrying medication when leaving home, and differentiating pills and tablets by color, form and size. Such items reveal that medication adherence is supported by ease of access to medication and by incorporation of medication into a daily routine, also a finding observed in previous studies (10,24) .
Finally, positive beliefs in regard to self-efficacy in treatment adherence were: understanding the prescription and how the medication is used. The negative belief in this category was forgetfulness.
These data corroborate findings in the literature (20)(21)(22)(23)(24)(25) . A similar study (19) identified most of the beliefs concerning DM medication adherence as being related to low adherence, highlighting the perception that there is no need to take medication when glycemia is under normal levels, concern over side effects, or due to complexity of the therapeutic scheme and dependence in following it.

Conclusion
This study's results show that adherence to oral antidiabetic agents is a complex behavior permeated by